Aims: To explore whether fluid resuscitation with normal saline or 4% albumin is associated with differential changes in routine clinical coagulation tests.
Design: Substudy from a large double-blind randomised controlled trial, the SAFE (Saline versus Albumin Fluid Evaluation) study.
Setting: Three general intensive care units.
Patients: Cohort of 687 critically ill patients.
Intervention: We randomly allocated patients to receive either 4% human albumin or normal saline for fluid resuscitation, and collected demographic and haematological data.
Methods and main results: Albumin was administered to 338 patients and saline to 349. At baseline, the two groups had similar mean activated partial thromboplastin time (APTT) of 37.2 s (albumin) v 39.1 s (saline); mean international normalised ratio (INR) of 1.38 v 1.34, and mean platelet count of 244 x 10(9)/L v 249 x 10(9)/L. After randomisation, during the first day of treatment, the APTT in the albumin group was prolonged by a mean of 2.7 s, but shortened slightly by a mean of -0.9 s in the saline group. The INR did not change in either group, while the platelet count decreased transiently in both groups. Using multivariate analysis of covariance to account for baseline coagulation status, albumin fluid resuscitation (P = 0.01) and a greater overall volume of resuscitation (P = 0.03) were independently associated with prolongation of APTT during the first day.
Conclusions: Administration of albumin or of larger fluid volumes is associated with a prolongation of APTT. In ICU patients, the choice and amount of resuscitation fluid may affect a routinely used coagulation test.