Objective: To test whether fluid resuscitation with normal saline or 4% albumin is associated with differential changes in acid-base status and serum electrolytes.
Design: Nested cohort study.
Setting: Three general intensive care units.
Patients: Six hundred and ninety-one critically ill patients.
Interventions: Randomization of patients to receive blinded solutions of either 4% human albumin or normal saline for fluid resuscitation.
Measurements and main results: Albumin was given to 339 patients and saline to 352. At baseline, both groups had a similar serum bicarbonate, albumin, and base excess levels. After randomization, bicarbonate and base excess increased significantly and similarly over time (p < .0001). On multivariate analysis, fluid resuscitation with albumin predicted a smaller increase in pH (p = .0051), bicarbonate (p = .034), and base excess (p = .015). The amount of fluid was an independent predictor of pH (p < .0001), serum chloride (p < .0001), calcium (p = .0001), bicarbonate (p = .0002), and base excess (p < .0001) on the first day of treatment. In patients who received >3 L of fluids in the first 24 hrs, albumin administration was associated with a significantly greater increase in serum chloride (p = .0026). Acute Physiology and Chronic Health Evaluation II score and the presence of sepsis also independently predicted changes in several electrolytes and acid-base variables.
Conclusions: When comparing albumin and saline, the choice and amount of resuscitation fluid are independent predictors of acid-base status and serum electrolytes. When large volumes are given, albumin administration leads to a higher chloride concentration. However, overall differences between the types of fluid are minor, whereas the volume of fluid administered is a much stronger predictor of such changes, which are also influenced by illness severity and the passage of time.