“A Comparison of Albumin and Saline for Fluid Resuscitation in the Intensive Care Unit” – The SAFE Study Investigators, The New England Journal of Medicine, May 27, 2004
One of the most powerful acute interventions in medicine today is fluid resuscitation. It is possible to instantly increase perfusion to all vital organs, saving unfathomable amounts of nutrient and oxygen-deprived tissue with adequate administration of fluids. The fluids that are administered are simple sterile water-based concoctions that do not carry the adverse effects or complications of blood or other medications. Therefore, initial responses to patients in extremis, those who require immediate and acute care, usually include obtaining vascular access and infusing large amounts of fluids.
Although there are many types of fluids, two major categories provide two physiologic approaches to fluid management. Crystalloid solutions are solutions of water and small molecules – small enough to leak out of capillaries and equilibrate with extracellular body fluid (body fluid that surrounds cells). Colloid solutions are solutions of water with large molecules – large enough to not travel through intact vascular membranes (they do not leak out of healthy blood vessels). Among other differing effects, colloid solutions tend to hold water inside the vessels better, but they can increase intravascular density, predisposing high-viscosity-low-flow states. Prior to this publication, small disputing studies had shown both benefits and lack of benefits from choosing one fluid type or the other. The SAFE investigators settled this dispute with a large (~7000 patients), double-blinded clinical trial that compared fluid management between 0.9% Normal Saline (NS, a crystalloid) and 4% Albumin (a colloid) in Intensive Care Unit (ICU) patients with the primary outcome being all cause mortality after 28 days.
After initial fluid resuscitation with Albumin or NS, there was no difference in mortality after 28 days in all patients admitted to the ICU. Patients given NS tended to receive more fluid initially resulting in a greater net positive fluid balance than those that received Albumin. However, even when broken down by common presenting illnesses to the ICU – sepsis, trauma or Acute Respiratory Distress Syndrome (ARDS), mortality was unchanged between Albumin and NS in each individual group. In fact, the only significant result this study produced was that NS administration produced lower mortality than Albumin in trauma patients with brain injury, a very small subgroup of all patients in this study.
Why We Do What We Do
We give NS. We give Albumin. The SAFE study showed that the two main fluids categories and the two most popular fluids are relatively safe. When administered to ICU patients – those least able to defend their body against physiologic challenges – neither fluid identified itself as inferior or superior. So was this study worthless because it did not identify a single solution and change or guide medical practice? I contend that strong, robust studies such as the SAFE study that do not find significant differences actually strengthen medical practice as a whole. Knowing these results provides doctors with both options of fluid when treating a patient. More importantly, doctors are able to use logical physiologic evidence, derived from basic scientific principles, to tailor their choice in fluids and management to the presenting patient.
Large studies such as SAFE are designed to adequately test and compare interventions. However, these studies are not intended to dictate prognosis for the patients. The data does not imply that one patient resuscitated with albumin will have the same outcome as if he had been treated with NS. Each patient is different, and giving the clinician the opportunity to apply his knowledge and training along with the guiding clinical evidence is how optimal care is provided.
Finfer S, Bellomo R, Boyce N, French J, Myburgh J, Norton R; SAFE Study Investigators. A comparison of albumin and saline for fluid resuscitation in the intensive care unit. N Engl J Med. 2004 May 27;350(22):2247-56. PubMed PMID: 15163774. http://www.ncbi.nlm.nih.gov/pubmed/15163774